Kernel Negative ε Dragging Linear Regression for Pattern Classification

Table 5

Accuracies (%) of different methods on the YaleB database.

Number of training samples per class	5	6	7	8
Our method	80.48	81.67	82.28	84.55
DLSR	76.05	78.48	79.33	77.54
CLSR	76.83	78.87	80.11	77.85
NNLS	68.98	73.39	73.38	75.14
-SVM	74.45	76.53	75.58	76.93
KNN	57.56	60.42	60.14	60.92
SRC	82.53	84.54	85.22	85.29
LRC	79.14	81.01	80.87	84.36
NDLR	77.02	79.40	80.25	78.73